Lyrebird · Realistic synthetic test databases, generated fully offline

See the output

Real source databases. The synthetic copies next to them.

Four public sample databases, one per engine. Open any table to see its source rows and the rows Lyrebird generated from its profile together, with the same measurements on both.

Databases
4
Tables
52
Synthetic rows
66,373
Broken foreign keys
0
75 foreign keys checked on the data

Generate them yourself

Download a profile and open it in Lyrebird Community to generate your own synthetic copy, with no source database. Community generates up to 5,000 rows per table and keeps complete reference data as it is. Each database page has the steps.

How these pages were made

  1. We loaded each public sample into its own engine: PostgreSQL, MySQL, SQL Server and Oracle.
  2. Lyrebird read each database once and wrote a profile, then the profile was sealed.
  3. Lyrebird generated SQL from the profile alone, and we loaded it into a second, empty database of the same engine.
  4. A script read both databases, measured every column the same way, checked every foreign key on the data, and saved the results.
  5. These pages are a snapshot of those results. No database runs behind this site, so queries are not available here.

What to look for: a value list such as countries or job titles can stay the same on purpose, and each table shows how much of it does. Generated names come from built-in lists, not from the source, so a common name can match one in the sample. Each column reports this.

Plan your enterprise evaluation

Let’s discuss
your test data requirements.

Share your database stack, target scale and deployment requirements.
We’ll help your team assess the right fit.

Request a demo info@subsetra.com